As the network ages and we have answers that are 7+ years old, we run into situations where changes in technology, etc. result in a once great answer becoming out-of-date, obsolete, or somewhat problematic. We don’t want to lose or delete these answers, due to legacy applications that still need the code but we need to figure out the right way to handle them.

This has been brought up and discussed many, many, many times over the years, and the community seems to suggest the following steps to update these answers:

  • Comment to the OP suggesting that their answer is out-of date and needs to be updated
  • Post a new answer with the updated information
  • Edit or suggest an edit (if you don’t have privileges) to correct the answer

There are pros and cons to each one of these, and I’m going to outline a few of them here:

Adding a Comment:

What if the user is no longer active on the site? They won’t get the notification to update and then it’s up to someone else to potentially edit the answer.

What if the user gets notification and disagrees that the answer is wrong, so they don’t edit it. Now we have a conflict between the original author and another user on the validity of an answer, how do we proceed?

Posting a New Answer:

This is a great solution to the above problem when the original author doesn’t want to edit their obsolete answer, but we run into the “sorting problem” where the highly-upvoted, accepted answer appears above an up-to-date answer. The hope is that eventually this new answer will gain enough upvotes to move up the list and will be easily visible to people with the same question.

Editing or Suggesting an Edit

This is where we have the biggest issue when it comes to improving out-of-date or obsolete answers. Editing posts is encouraged on all sites and it’s even a privilege that is earned. The help center explicitly states

When should I edit posts?

.... to correct minor mistakes or add addendums / updates as the post ages

Users are advised to make these edits, yet this seems to be a stumbling block.

Why, you ask, do I see it as a stumbling block?

Several reasons...

First, when a user doesn’t have full edit privileges it goes to the review queue where these edits can be rejected. If the edit is substantial enough, there is even a rejection reason provided:

This edit deviates from the original intent of the post. Even edits that must make drastic changes should strive to preserve the goals of the post's owner.

With the rejection we lose the proposed fix and are back at the start, with an answer that needs a bit of tending to.

If a user has full edit privileges, they can make a sizable edit without review to bring it up-to-date but if there are code changes users are told not to touch the code.

As you see, we have a disconnect. Users are told to edit and we have posts that need to be edited. But no one wants or is able to edit them, because they might alter code which is a big no-no when the answer is not a community-wiki.

What do we do? What can we do to fix this?

We’re taking a good hard look at how we should handle answers that were once great but need some gardening to keep them that way. Lots of things have been suggested over the years, but so far nothing has been implemented:

Before we do anything we want some feedback and a discussion from the community on how we can encourage these edits to these answers (we disagree internally as well).

Would you be more likely to edit them into something no longer obsolete, if they were made community wikis? Should we add a post notice that it’s obsolete or out-of-date? Figure out some way to give rep to the user who makes a substantial edit? Something else that we haven’t thought of?

  • 50
    Not sticking the accepted answer to the question would be a start. Then if and when a new answer gets more upvotes it will rise above the old answer.
    – ChrisF Mod
    Jul 31 '15 at 21:05
  • 16
    What about having something similar to "duplicate" but "obsolete" that redirects to a new question with a proper solution to the problem? This way the former question remains valid as it was and the new one is the new, good way to do things. Jul 31 '15 at 21:29
  • 4
    @fedorqui These might not be separate questions though. It might be a newer answer to the same question.
    – Taryn StaffMod
    Jul 31 '15 at 21:36
  • 84
    One major issue is people religiously close as duplicate any new question, but then visitors get angry when the 7 year old question has obsolete answers.
    – nhinkle
    Jul 31 '15 at 21:45
  • 52
    Another big problem with editing the answer is that the original answer is good for legacy systems.
    – msh210
    Jul 31 '15 at 22:05
  • 17
    We run into this on RPG.se often enough, with games that have changing/updated/errata'd rules. We do two things: 1) ask active users to update their answers, 2) edit old questions to rescope them to exclude whatever new thing made them/their answers obsolete to make room for new non-duplicate questions. (1) is unreliable as noted here but works sometimes, and (2) doesn't always apply; but together it catches a large fraction of the problem. That said, it's not really satisfactory either. Jul 31 '15 at 22:12
  • 45
    I'm having trouble seeing how the "obsolete" posts will be objectively identified. There are new questions posted every day about obsolete technologies on Stack Overflow. Should every Windows related answer that doesn't apply to Windows 10 be considered obsolete now?
    – Ross Ridge
    Jul 31 '15 at 23:00
  • 6
    And 80% of c++ questions (hello, Turbo C++ from 1990 in Indian locale mode) Jul 31 '15 at 23:13
  • 9
    I agree with msh210 & Ross Ridge - "Legacy" systems exist and many are actually production systems - Knowing the fix for a 7 year old problem is valuable information. Perhaps a solution to is to introduce a "Search by Question Date" option, so users can find answers to things they know are new problems. I use the google version of that feature all the time to find current information. People that wish to can still update their answers to reflect current information, in response to down votes (I've done that).
    – Seth
    Jul 31 '15 at 23:48
  • 7
    As @Earthliŋ said, you might be trying to treat a symptom instead of curing the disease. Stack Overflow is in decline, and broken windows are just a manifestation of that. Aug 1 '15 at 6:00
  • 4
    Do we also need to change the culture that emphasizes rejecting changes that "clearly conflict with the author's intent" or "putting words in the OP's mouth/text" or "correcting/adding code to existing answers rather than posting a new answer"? The technical changes are ease compared to culture changes.
    – user213963
    Aug 1 '15 at 13:14
  • 14
    On the one hand We don’t want to lose or delete these answers, due to legacy applications that still need the code but we need to figure out the right way to handle them and on the other hand you are asking to edit those answers? That's a contradiction to itself. As soon as the answer is edited, the old original answer is lost as it would be when it is deleted.
    – baao
    Aug 1 '15 at 14:46
  • 8
    I see this mentioned in other comments, but I don't see how we can call a question or answers obsolete. Maybe we need to do a better job of identifying specific versions and technologies in questions - take a C++ or Java question. One asked 7 years ago about C++03 and Java SE 6 would have different answers today, if you were using C++11 and Java 8 since new capabilities have been introduced. But that doesn't mean that the answer isn't valid anymore and shouldn't be weighted less since it's still incredibly useful to someone using those technologies. Aug 1 '15 at 16:55
  • 18
    From my own history, I have an answer to a question that specifically referenced a particular version of Eclipse and asked how to do something that was not possible. It's gotten comments pointing out that it was "wrong" and even downvotes. However, to anyone still using Eclipse 3.4.2, this is the correct answer. I ended up adding commentary to my answer (which appeared to stem the down votes), but I shouldn't have to do that. Questions may be tied to things that change over time. That's OK Aug 1 '15 at 16:58
  • 27
    PLEASE save obsolete answers, and link them. And encourage version numbers in questions. This is very useful, especially in technology, when there are different solutions for different API levels. In Mechanics, there is a similar value, etc. And, tackling a physiques problem from Quantum or Classical physics could be very helpful as well. Aug 3 '15 at 16:42

37 Answers 37


My vote is for Obsolete answer.

As you have stated, the sorting would stop the new answer from being on top, which means the obsolete info is what the curious guy with the same question will face. This is a big con of adding a new answer, as I've seen lots of better newer answers not getting more votes than the old answer. Votes on answers marked as obsolete wouldn't affect the sorting order.

Now, this can either be a mod-flag review case, or a privilege. I don't think treating these votes like "moderator notices" or "flag reviews" is really helpful, as mods in a site as humongous as SO will get tired of reviewing them soon. So, casting a vote as a privilege - a "review-grade" privilege - would do the job.

But please:

  1. Can this be a higher level privilege than close/reopen votes? I'm imagining at least a 5k on betas and 30k on the graduated sites. (Note that the number isn't important, as long as unjustified votes are at a minimum)
  2. Make the template message sound more human and more serious.
  3. Make the votes reversible. i.e. if an answer with this mark is edited, or a VANO (vote-as-not-obsolete) is cast, it "enters the queue". Much like the "reopen queue".

Here's why: Wikipedia sometimes gives warnings about its articles:

I for one, don't think they're that useful. Why? I've seen them so much in the articles I quote a lot, use a lot and cite a lot, that they're not really indicating anything serious to me anymore. This feature would be useless if the messages end up being around in almost every post from back then, or they are seen a lot in posts that don't really need them.

That being said,

Maybe they should get their review queue on the trilogy (SO, SF & SU), since I imagine it'd be frequent enough not to be like this:

This could also work well in sites as big as math.SE, or even physics. But, Newer sites, and the ones with less traffic won't really need it IMO. Those posts could simply bump into the VLQ queue, with new options: "Looks OK" and "mark as obsolete".

Also, to encourage people to write answers to these old posts, you can exclude "obsolete" answers from counting towards the question's "answered" status. Which means, if the answers to a question are marked obsolete, and there are no other upvoted answers, the question would appear in the "unanswered" questions' list. ("need answers" is the new-nav version) This would also bring attention to an updated answer once it's written or needs to be, which is what we aim at.

The other ways to encourage updating info, that are editing or commenting would hardly get as much traction. The closest would be to see the post active, but that only would happen if the answerer agrees that their answer should be changed, or the answer is community-wiki (this itself makes it less-noticeable), or the edit somehow managed to get approved and not rejected with a seemingly legitimate reason.

  • 12
    How does this encourage editing of these answers? There will be a post notice but we will still run into possible issues on who is going to do the edit? What if the author disagrees that it's out of date? The goal is to tend to these answers to bring them in line with current technology, users wouldn't be hesitant to making these edits if it were a wiki.
    – Taryn StaffMod
    Jul 31 '15 at 21:56
  • 6
    This vote should stop the votes of the old answer from making it appear as the top answer. This way, a newer answer could appear on top with fewer upvotes. In other words, I'm trying to eliminate the cons of adding a new answer. Edited the answer, thanks for pointing out.
    – M.A.R.
    Jul 31 '15 at 21:59
  • 10
    @bluefeet Why is the goal to bring these answers in line with current affairs? SE want's the best answers by any measure, and "oldest" shouldn't automatically merit rescuing because "oldest" has no correlation with "best." Replacing or adding to old material that's still relevant to old systems is an issue too. In the long tail, "How do I configure NAT under Windows" shouldn't result in one answer that details how to do it under 100 Windows versions, when the question was originally just about Win95. Jul 31 '15 at 22:27
  • 1
    @SevenSidedDie We want a repository of the best answers, if there are problematic answers (for example) due to tech changes, there is always a lot of discussion about what to do with them. We don't want to delete, ideally we want a newer answer to be seen or the older answer to be improved to incorporate those changes. I'm not saying oldest is bad, but there are cases when these need to be tended to.
    – Taryn StaffMod
    Jul 31 '15 at 22:33
  • 9
    @bluefeet So there's a distinction that I think's being missed then: there are obsolete answers, and then there are harmful answers. The second need remediation, but I don't think the first would necessarily be solved by the same remedy. Jul 31 '15 at 22:36
  • 4
    Well I completely agree with @SevenSidedDie. It's typically useful to have obsolete information around, but I can't think for one minute that the same holds for dangerous content. Blame my chemistry background for my example: If an answer in chem.SE is invalidated because of the newer IUPAC rules of nomenclature, then a new answer is more preferable, stating separately how the new rules hold. But if the answer is about handling a dangerous chemical, which is later proved to be a wrong way of handling, then I hardly think the consensus would be "leave the answer alone", let alone editing.
    – M.A.R.
    Jul 31 '15 at 22:45
  • 5
    @inɒzɘmɒЯ.A.M I may have put more emphasis on "harmful" than necessary. On further thought, I'm thinking that the ways answers can become obsolete is diverse enough that we might all be picturing and therefore arguing about different things. That might be factoring into the internal dissent mentioned in the question too. There may not be a one-size-fits-all remedy for obsolete answers; what bluefeet linked to there is certainly a different kind of obsolete than the practical examples I was thinking of, and measuring this meta Q against. Jul 31 '15 at 22:49
  • 2
    @Zeks voting has always meant the ability to affect a post in some way. But I agree that there could be flag reason for it which bumps it into the queue. What I worry though is that opening this door would flood up the queue, leaving the answers that need urgent attention unnoticed for at least a while. I might be wrong though. I hope I am. ;)
    – M.A.R.
    Jul 31 '15 at 23:25
  • 1
    For the record, in theory anyone can remove those Wikipedia notices at any time if they are not applicable. They are not "official" any more than the rest of the article's content. In practice, of course, this may require persistence to "stick.".
    – Kevin
    Aug 1 '15 at 0:19
  • 2
    I think this can work. In fact, I have an answer myself on SO where I added a blockquote at the top, that mentions how my answer is downright wrong and that another answer on the same question is much better than mine. If you look at the timeline, you can notice that @JoeConlin's answer started getting more attention/upvotes after the edit. Maybe SE staff can validate this with more analytics on the post, but IIRC, my answer was the most upvoted before I added the quote, and there were 10k+ views.
    – Dan
    Aug 1 '15 at 9:04
  • 1
    This could also work well in sites as big as math.SE, or even physics. But, Newer sites, and the ones with less traffic won't really need it IMO. – I do not see anything wrong with a review queue that is hardly ever used. On some sites, getting a Reviewer badge for late answers is near impossible. Also note that Math is bigger than SU and SF right now (but then answers rarely, if ever, become obsolete there).
    – Wrzlprmft
    Aug 1 '15 at 11:15
  • 2
    If a Wikipedia warning is not relevant any more (especially a two years old one as in your example), you should click "edit" and remove it.
    – Nemo
    Aug 1 '15 at 13:22
  • 2
    Perhaps add an open "reason" when obsolete-voting, so that the notice says "This question is voted obsolete becasue (a) The foo no longer bars the baz since version 1.8, (b) Trying to blah the blah will cause your prod server to vomit lava". This will make the notice more useful. Aug 1 '15 at 17:25
  • 2
    I think the main benefit of this would not be to encourage edits to obsolete answers as much as it would be to encourage new answers which are more up to date, and to direct the attention of more users to those answers. If I'm searching for an answer to a problem and I see a question with a highly upvoted answer, I might not read all the other answers. If, however, that upvoted answer was marked with a notice saying it is obsolete, I'd be more likely to look below it for newer answers which might not be as highly voted (yet).
    – Ajedi32
    Aug 3 '15 at 20:15
  • 7
    Also, to clarify, answers shouldn't just be marked as "obsolete". Rather, the notice should state under which circumstances the answer is/was applicable. (E.g. Caution: This answer may be outdated because it only works with Windows XP. or This answer may be outdated because it no longer works in versions 2+ of <some software>)
    – Ajedi32
    Aug 3 '15 at 20:18

The proper solution to this problem was suggested 5 years ago and is the #18 overall highest voted answer on all of MSE (higher if you don't include Jon Skeet Facts answers):

If we provide this sorting formula, the newer, correct answer can get upvotes, which can make the newest, correct answer leap ahead without having the years of time to attract all the upvotes. At the very least having "Best" could be offered as a tab while we see how effective it is.

@Shog points out in the comments that this might not be the right thing to do in the case of answers that are NOT obsolete and encourages plagiarism and pile-on answers. Perhaps the right thing to do is leave "Votes" as the default tab sort for most questions, but use "Best" sorting for questions that meet a certain criteria?

  • 9
    Yes, yes, yes! I've been looking for this tab so often. Why is a -1 accepted answer still at the top, when the OP hasn't visited the site for 2 years and other, 13 or 22 score answers are below it...
    – Earthliŋ
    Jul 31 '15 at 23:52
  • 41
    The big danger of this is that most answers are not obsolete, and favoring new answers to old questions that are already well-answered both fails to serve readers and tacitly encourages pile-on answers (or even straight-up plagiarism). This can be avoided perhaps (vetting period before new answers qualify?) but not a silver bullet. It might make more sense to apply this only to questions where some indicator exists that obsolete answers are a problem.
    – Shog9
    Aug 1 '15 at 0:15
  • 20
    @Shog9 As I understand it, it favours neither new nor old answers, as eliminating “fastest gun” voting bias is exactly its raison d'être. If that sorting algorithm favours new answers, it is implemented wrong. However, cautiously implementing it only on questions that have been decisively flagged as having obsolete answers may be very useful (and a nice data source for proving its effect). Aug 1 '15 at 1:41
  • 3
    I think the effects of duplicate answers could be managed via either flags or downvoting, which is the traditional way to discourage duplicate answers today anyways. Aug 1 '15 at 1:49
  • 5
    @PearsonArtPhoto I would point out that not enough people down vote. Also flagging a "this answer is a duplicate of this other one and is wholly a subset of the information contained within" is a very hit or miss proposition on many SE sites with many moderators. The general SE populace has been taught that flagging is a last resort because unless it is extremely clear cut and doesn't require moderators to evaluate the content of the answer, the flag will be declined.
    – user213963
    Aug 1 '15 at 2:37
  • 2
    @SevenSidedDie I think the point Shog9 may have been making is that, if the existing answers on a question are good, those answers should be favored over new ones. So an algorithm that removes the effect of age would be favoring new answers more than it should be. The ideal sorting algorithm would make new answers actively prove their merits (through upvotes) before they make it to the top spots, but without "hiding" those new answers from potential voters.
    – David Z
    Aug 1 '15 at 4:35
  • 9
    @DavidZ Those would be valid criticisms of a different algorithm that did promote new ones before they proved themselves with any votes. I suspect that reading the explanation of the algorithm would clarify things. Aug 1 '15 at 4:37
  • 2
    What if this 5 years old answer on meta is obsolete ? ;) Aug 3 '15 at 12:51
  • 1
    @AleksandrDubinsky Can we actually see some results before we start criticizing the algorithm? I want to see what sorting actually occurs first before we try to tweak it.
    – durron597
    Aug 3 '15 at 21:35
  • 1
    @AleksandrDubinsky We're only talking about questions that get tens or hundreds of votes.
    – durron597
    Aug 3 '15 at 21:44
  • 7
    @Shog9, The algorithm doesn't give an advantage to newer answers at all. What it does is that it puts all answers on equal footing, and thus so-called "encouraging plagiarism" is unlikely because plagiarized answers would get their fair share of downvotes. Without downvotes, the plagiarized answer would float to the top and gain maximum visibility, and thus the chances of it being downvoted increases as time passes. The algorithm is good and indeed worth a try.
    – Pacerier
    Aug 4 '15 at 11:03
  • 1
    There's another beneficial side-effect of the algorithm. Some answers are not obsolete, but are one-or-two-liners which were cemented in top position due to fgitw problem. While these answers are not inaccurate in content, they lack substance. The new algorithm will encourage people to post better answers with clearer elaborations. If the new answers are better than the one-liners, they would replace it as top position. On the other hand, if the one-liners are indeed better than the new answers, they would remain as top position.
    – Pacerier
    Aug 4 '15 at 11:19
  • 3
    @Shoq9 and 28 voters haven't understood the algorithm but are influencing the debate. That's dangerous.
    – FooBar
    Aug 4 '15 at 18:16
  • 1
    @Shog9: You misunderstood my comment on the linked answer. The equation given does not favor new or old answers, at all. My suggestion was to add a small bonus to the equation, say, 1 / numberOfMinutesSincePosted. This would push new answers to the top, but only for a few minutes. If the question is new, this would prevent the first answer from hogging all the upvotes. If the question is old, only people who hit 'f5' on the active-questions page would ever see it at top (where they can conveniently flag or upvote); people who are actually looking for that question would not.
    – BlueRaja
    Aug 4 '15 at 22:11
  • 4
    @Shog9: The problem is that most people only read the first answer, so when a new answer is posted, the first answer to get any upvote (typically the first answer posted) will most likely be the one to get most of the upvotes. This results in the simpler questions, the kind that get 5 answers in the first 10 minutes, to be a race to see who can answer first. The shorter, lower quality answer then tend to get all the upvotes, while the higher-quality answer posted a few minutes later (because it's longer) gets less. Hence the term "fastest gun in the west".
    – BlueRaja
    Aug 6 '15 at 16:02

I'm just throwing this idea out there for an alternate solution to consider.

I love the "Linked Questions" section. Why not expand that a bit?

In most cases if I'm looking for another version to a question, the first thing I check is the "Linked Questions" and "Related Questions" sections. Usually I'll find what I'm looking for, but it would be nice if there was a dedicated section for "Alternate Versions of this Question", especially in cases where there is a really long list of linked or related questions.

To me, it makes more sense to post new questions with new answers/information if referring to a different version (or even slightly altered requirements or attributes), rather than curate an old question that uses old/obsolete technology with all new answers and information. The important part is to have them linked in a way that is easy for users to find.

We already have version-specific tags, and we do allow for questions along the line of "I already saw Question 123, however it is using an older technology/framework and I am looking for an answer using New Technology/Framework", so it seems natural to continue this route.

The initial implementation is already there - users just add a comment to the question or accepted answer with a link pointing to the newer/older version of the question, and it shows up in the "Linked Questions" list.

If you wanted to get fancier, you could allow entitled users (gold-badge users?) to tag the link as an alternate version of the question to show up in a new section.

And it would be scalable in the long run too. It gets hard to curate posts for all possible versions of a question, and you've already pointed out problems related to the voting system or users that are no longer active. This solution would keep each version separate, but linked.

enter image description here

Anyway, it was just an idea I thought I'd throw out there to consider, as all options so far are focused on taking action on the original question itself rather than looking at other SE features.

As for the second part of your question, here's my personal answers :

Would you be more likely to edit them into something no longer obsolete, if they were made community wikis?

No, I'm lazy and if people want to know a newer version they can Google or ask a new question

Should we add a post notice that its obsolete or out-of-date?

No, as someone who uses this site frequently I would find this very misleading and confusing, especially if the information actually was current despite its age. I know enough to look at the date of information posted when evaluating it.

And as someone who cares about the site, if this post notice was added by anything other than an automated process based on something like age, I would find the maintenance of such a thing daunting.

Figure out some way to give rep to the user who makes a substantial edit?

The type of people who can make substantial edits to address other versions are likely those who have been around long enough to not care about rep. Personally, this would be no incentive for me.

Something else that we haven't thought of?

The only thing I can think of that would make me want to write an updated answer to an older question would be someone asking for an updated answer to help them out.

If users could have a button to say "I think this information is outdated and would like to request a more current answer", I would attempt to help them if I could. We already have a system for that (bounties), but think it is too broad and requires too much personal investment to really be viable for promoting newer answers on older questions in a wide scale.

  • I don't see why "the type of people who can make substantial edits to address other versions" would be enough different from "the type of people to write really good answers" to justify rewarding lots of rep to the latter but nothing special to the former. A really good update edit requires a certain amount of expertise in not trampling on the original, but not nearly as much as it requires in writing the updated details, and I would not be surprised if there are a good many users with <2k that can manage it. Aug 1 '15 at 19:09
  • 1
    @NathanTuggy Yeah, I guess I meant to say the type of people who know we want obsolete answers edited with updated information, and capable of providing a great answer, are typically those who already have a lot of rep or don't care about it :)
    – Rachel
    Aug 1 '15 at 19:34
  • 4
    +1 to encouraging links. Whether the approaches to linking should be expanded, I can't say; my mood, in response to the complex set of questions in the opening post, is keep it suitably simple. Aug 2 '15 at 5:11
  • In some cases adding a version tag where there isn't one would help. Maybe an answer can be edited to add a statement indicating that it was for a specific version.
    – Elin
    Aug 2 '15 at 18:40
  • "Should we add a post notice that its obsolete or out-of-date?" " I would find this very misleading and confusing, especially if the information actually was current despite its age." How would this be misleading? It is only added when it is true. Aug 2 '15 at 22:20
  • I really like your idea to have a section to up-to-date technologies. But that brings a huge problem: Tags. Yeah, tags will be a pain. Some questions only have php while others have php-5. And now? How do you separate both? That will be the tricky bit. Also, there would be no indication that the content is out-of-date. Aug 4 '15 at 8:34
  • 1
    @IsmaelMiguel I was thinking whoever is tagging the posts as alternate versions of the same thing should also take care of tagging them. I think the community is pretty good about retagging older versions already though... I know I've forgotten to tag my posts with a version # on occasion when asking about older technologies, and someone always seems to edit my post to add it. SE could even consider if it would be worth it to take the tag system a bit further and extend it to links.
    – Rachel
    Aug 4 '15 at 13:33
  • Extending to links would almost be a good idea. Yeah, there's tons of people good at retagging. Perhaps your idea isn't as clunky as I had in mind. But still, it is a lot of work. Aug 4 '15 at 14:15

As I have said before, please bring out the stats and/or cases. How many of the answers do you consider obsolete? How much is it a problem, actually? You are encouraging fiddling with the basic mechanism of Stack Exchange without the facts, blindly, as it seems.

Would you please:

  1. Create a Stack Exchange Data Explorer (SEDE) query or conduct a survey of a random sample of Q&As (if you don't want to scour through the whole SO question base).
  2. Post the results of the query.
  3. Provide sample answers that you consider obsolete.
  4. Explain what makes you think that existing mechanisms will not work to fix the problem.

Now back to the substance of the question, which is the accumulation of technically inaccurate and obsolete cruft:

  1. I don't deny there is such a problem. Developers are pushing forward backward and sideward incompatible new versions of popular frameworks and what used to be the state of the art solution just a month ago is no longer correct now.

  2. Accumulation of cruft is caused by lack of curation. Wikipedia resolves this by a cadre of editors jealously watching their own articles and a thick layer of bureaucracy resolving conflicts. Stack Exchange was founded on the premise that voting lets the nuggets of truth in the "long tail" percolate to the top and pieces of ... well, other stuff drop to the bottom. For this to work, one needs minimum 30 to 40 (ideally 100 to 200!) technically proficient professionals who can comment on unsound solutions, moderate things, vote, post their own answers, curate outdated but popular answers etc. Why the threshold of 30 to 40? It's when the law of large numbers kicks in - below that you have to rely on a handful of people who aren't always available.

  3. For older sites in the Stack Exchange network and even some tags on Stack Overflow, this minimum is no longer available.

  4. We are asked to suggest technical solutions to a peopleware problem. This is a great approach, isn't it? What's more, we are asked to consider changing the fundamentals of Stack Exchange. There can't be any unintended consequences, right?

  5. Having objective data and clear understanding of how the system works could have prevented us from getting to #4 with a "great approach". We have to talk about the root cause, not about the vague symptoms of malaise.

  • 6
    After all, we require posters on main sites to show an MWE and proof of own research. The same standard should apply to MSE. Aug 1 '15 at 8:34
  • please, can you link to where those things were previously written by you? Also I suggested an edition to this answer – primarily with attention to the word blind – "hopefully not losing the essence …". Aug 2 '15 at 3:38
  • 2
    Rapidly-evolving technologies generate large numbers of obsolete answers. meteor is a good example if one is familiar with Meteor.js. Aug 3 '15 at 8:27
  • @DanDascalescu - thanks for the -, there are only 15 active experts in the meteor tag, what do you expect from them? Aug 3 '15 at 8:42
  • I don't expect anything of them; it's a volunteer effort. I hope that SE implements more effective ways to bring updated answers to attention, and more importantly, changes the status quo that prevents updates to begin with (see Michael's answer). Aug 3 '15 at 8:57
  • 3
    I agree. Most old questions that need new answers don't have "obsolete" answers. They have bad answers. I try to write one when I'm able, but the issue is that it's generally a rewardless experience. SO users only vote on new questions. Aug 3 '15 at 21:11
  • 5
    To deem an asnwer "obsolete" or "requires revisit" is a human criterion, that can't really be automated. And unless real people go in and add a new answer (or update/comment on the ones extant), nothing will happen. Maybe flagging as "please consider adding a new answer" and bubbling up into "new" question status would help. Allow sorting answers by accepted/not, votes, or from newest to oldest could be useful for archeologists...
    – vonbrand
    Aug 4 '15 at 1:47
  • Almost the entire Swift language section on SO falls under the "obsolete" part now. A good 90%+ of the answers are directly wrong in Swift 2.0, which is a big problem. Aug 4 '15 at 5:40
  • @ClausJørgensen - only 5-9 folks are active in the swift2 tag. Aug 4 '15 at 5:58
  • 1
    @Pacerier - content has to be curated by somebody. If the number of curators stays the same (or even declines, like for many top tags in SO), and content explodes, or just grows linearly, no technical gimmicks are going to help. Here are the stats for the php tag. Aug 4 '15 at 11:35
  • 2
    @DeerHunter I get a feeling that after AMAgeddon at Reddit StackExchange team members have been given orders to engage the community by posting discussion seeking questions every other day. Wow, really? This was posted because internally we feel it's a problem that is happening on the site and before proposing any type of fix to the network at large, we want feedback on it. That's the way the network has always worked.
    – Taryn StaffMod
    Aug 4 '15 at 12:47
  • 5
    I think some examples of what would be considered "Obsolete" are very important. I am unclear how broad or narrow the scope would be, and feel if this route is chosen this has to be defined better. Would every single .Net tag except the most recent version be considered obsolete? Would a question and answer that targets .Net 3.5 be considered obsolete just because .Net 5.0 does it differently now? Or would the scope be extremely narrow and only count posts that actually no longer work or are no longer a problem anymore due to technology changes?
    – Rachel
    Aug 4 '15 at 13:40
  • 2
    @Rachel - agreed, we are treading on dangerously subjective ground. Aug 4 '15 at 13:49
  • 2
    @Pacerier However obsolete that answer was, your edit is one that completely changes the intent of the answer. This shouldn't be done like this. Leave an editor's note or something in such cases, and maybe add your own answer, but don't edit like that.
    – Bruno
    Aug 6 '15 at 11:58
  • 1
    @Pacerier I'd rather see each text attributed to its rightful author (and, no, expecting readers to go through the revision history to determine authorship isn't realistic). You can't anticipate whether the other author will edit or not. Your example is a different problem: it looks like the other answerer took substantial inspiration on yours to improve his own answer (although to be fair his link seemed to point to "MaxExactCount", so I'm not sure that's "cheating"). Having two similar answers now isn't a catastrophe.
    – Bruno
    Aug 12 '15 at 10:43

If a question is specifically about change in a technology, it isn't a duplicate.

If an answer to a given question has been made obsolete by a new version of a technology or regulation, you can ask and self-answer a new question that's specifically about the new version.

In the question: Avoid closure as duplicate by including the version number of the relevant change in the new question's title, citing the old question, and ideally citing the particular change that made the old question's answers not apply. It might look similar to the following:

The solution described in question "How do I do whatever?" works for Product 5.3, but someMethod() was deprecated in Product 5.6.

The link in this notice will put the new question into the "Linked" section of the old question. An example of a question using this m.o. is "Can I change my profile picture using the API version 2?" on Stack Apps.

In the answer: Stack Exchange uses the same Creative Commons license as Wikipedia. So feel free to adapt someone else's answer to the old question into your self-answer to the new question, so long as you cite that answer and its author.

  • 2
    The only problem is that you now have 2 obsolete questions and answers. And then you will have one more obsolete question and answer... And one more, and another one, and another, another, more and more and more.... This solves the problem for some technologies but not all. One language that will be a pain is Javascript. Each browser changes stuff around and new things come out and ES6 is there and we have a huge coding salad. Aug 4 '15 at 8:38
  • 1
    @IsmaelMiguel And the "Linked" section is the dressing that holds the salad together. Aug 4 '15 at 17:05
  • 1
    And quite a salad. It's like if you were trying to eat a piece of tomato in it, and the whole salad was linked and you have to eat it all at once because you can't separate the tomato from the letuce. I'm not saying your solution is bad, I'm just trying to show that it isn't optimal. It's not bad, just not optimal. Aug 4 '15 at 17:07

This started out as a comment in response to this one, but got too big.

For context:

Do we also need to change the culture that emphasizes rejecting changes that "clearly conflict with the author's intent" or "putting words in the OP's mouth/text" or "correcting/adding code to existing answers rather than posting a new answer"? The technical changes are ease compared to culture changes. – MichaelT

@MichaelT Yes I believe that is a large part of the problem. But the question is how do we fix that. – bluefeet♦

There is a culture that the answers are "owned" by the person who wrote them. That changing the text, no matter how old, should be rejected.

There is also a culture in the ranks of the moderators that every bit of information should be preserved. That they are not to do more than squash spammers and decline flags and maintain the status quo.

The problem here is that the status quo is an ever accumulating pile of crap with an ever decreasing signal to noise ratio. While the suggestions here are that users are to be given tools to either hide, reorder, or otherwise work around this, it is also clear from voting stats that this is a losing battle (far more upvotes on everything - the conception that the up vote is akin to the 👍 on Facebook or other social media sites).

There is a cultural problem and a technological problem and you need to fix both of them. they are not able to be separated. In A Group Is Its Own Worst Enemy Clay Shirky talks about an old lesson on LambdaMOO (note: any SE employees who are unfamiliar with this individual - check your board of directors - there's a reason he's there).

There's a great document called "LambdaMOO Takes a New Direction," which is about the wizards of LambdaMOO, Pavel Curtis's Xerox PARC experiment in building a MUD world. And one day the wizards of LambdaMOO announced "We've gotten this system up and running, and all these interesting social effects are happening. Henceforth we wizards will only be involved in technological issues. We're not going to get involved in any of that social stuff."

And then, I think about 18 months later -- I don't remember the exact gap of time -- they come back. The wizards come back, extremely cranky. And they say: "What we have learned from you whining users is that we can't do what we said we would do. We cannot separate the technological aspects from the social aspects of running a virtual world.

The point I'm making here is that beyond giving us (the community - not the diamond mods) the tools (long overdue) to work on moderating the site you need to step in and take active steps in getting the culture to be one that does not reward or preserve crap (obsolete answers are just part of the symptom).

It is not sufficient to just do a cosmetic reordering of the answers. You need to change the culture and attitude of those who say "when you see something wrong, down vote and continue on - trust the community as a whole to sort good material out form bad" because you know that this doesn't work.

I'm not saying that the tools of obsolete or reorder won't help - but you need to think bigger and be actively involved with the culture of the moderators, the 10k+ users, the 3k+ users, and the 15+ users. You need to work on changing the perception of a +1 being harmless and equivalent to a +like. And that may mean some bigger changes or getting your hands dirty in the community again.

  • While I greatly agree with you in general, I don't agree that we should be substantially editing other people's posts. This can be quite problematic, for the reasons outlined at meta.stackoverflow.com/a/293969/3153792
    – corsair992
    Aug 1 '15 at 20:47
  • 4
    @corsair992 Did you link to the wrong thread? On the contrary, meta.stackoverflow.com/a/293969 is a prime example of a case where edits were sorely needed. Aug 1 '15 at 23:42
  • @Gilles: I linked specifically to Bruno's answer in the thread. Perhaps the breadcrumb did not load for you for some reason, or you ignored or scrolled past it?
    – corsair992
    Aug 2 '15 at 8:28
  • 1
    @corsair992 Bruno writes “I would edit the answer saying something at the top…” He argues against modifying code (which I don't agree with, but that's not the point of contention here), but he doesn't go as far as arguing against editing the answer — transposed here, that would mean adding a sentence at the top of the answer to say something like “this answer only applies to version 3”. Aug 2 '15 at 11:39
  • 1
    @Gilles: I have no problem with adding a small editor's note at the top of an answer, that does not contain any code and is clearly not attributed to the author, as Bruno suggested. Nor am I even opposed to correcting obvious mistakes in the code. Anything further than that would be a mistake in my opinion though. I think we are in agreement here.
    – corsair992
    Aug 2 '15 at 11:50
  • 8
    "The problem here is that the status quo is an ever accumulating pile of crap with an ever decreasing signal to noise ratio." - exactly. Blind rule following has been pervasive on SE. We lack something in the spirit of Wikipedia's IAR. Aug 3 '15 at 8:30
  • 2
    @Gilles to clarify my answer linked in these comments, I'm generally against editing someone else's answers in a way that puts words in the other author's mouth (and expecting the casual reader to go through the edit history is just too much). However, I think that leaving a clear editor's note (something that's clearly an addition to the initial answer), for example in bold/italics or, if there was a template, some sort of wikipedia notice box would be perfectly acceptable. Note that that question was more about answers with security issues, the notion of "obsolete" is slightly different.
    – Bruno
    Aug 4 '15 at 16:22
  • 2
    The idea of the site at the very beginning was for posts to be wiki-like and encourage edits, minor or significant. I was disappointed with how quickly that spirit was lost, with people being very defensive of their answers even when many experts wanted to improve them. It would be great if we could turn that around.
    – Jeremy
    Aug 4 '15 at 17:19
  • 3
    @JeremyBanks That's never how I felt the site was. That wiki-like idea just doesn't work when answers have a main author name against them, and when there's a gaming system that is very much individualistic.
    – Bruno
    Aug 5 '15 at 13:18

When showing answers, take into account the age of the answer when deciding which answer to show first. If there is an answer that is only a few months old, and has 25 votes, compared to an answer 6 years old with 100, I suspect the answer with 25 votes that is newer is likely to be a good answer.

At the very least, an option for a newly voted question should be available, in addition to the one with the overall highest votes.

What this will do is allow users to post a new answer if there is a question that is very obsolete. This new answer that has been highly voted will show up prominently. This will encourage users to provide new answers to these old questions, knowing they still have a chance to get some traction from said question.

  • 16
    Formalized, this would seem to be the Wilson score solution. Jul 31 '15 at 23:21
  • 2
    What about “votes per view by voter”, so if a new answers has got +5 votes when viewed by 50 people that can vote and do often vote, it goes above a old answer that has got +10 votes when viewed by 500 people that can vote and do often vote. Aug 1 '15 at 13:50
  • That's an interesting alternative, but I don't know if that would work for existing questions, it might require another addition to the DB to make it work correctly. Aug 1 '15 at 13:52
  • @NathanTuggy Are you sure about that? I think PearsonArtPhoto is suggesting an implementation that uses age directly as part of the calculation. E.g., rankscore=(upvotes-downvotes)/age. Aug 2 '15 at 2:40
  • @RyneEverett: The rough sketch of the general algorithm seemed vague enough to me to be satisfied by the solid statistics of Wilson, but you're right that it does suggest using age specifically. Aug 2 '15 at 4:18
  • roughly equivalent to another answer about different ranking Feb 7 '16 at 2:14
  • @IanRingrose Since all answers are visible when the page loads (that's a good thing), how would you determine how many people viewed, or read, any particular answer? Dec 20 '19 at 3:35

I've had some recent experience with this.

How to add multiple font files for the same font?

The accepted answer was invalidated several years ago by the spec, but no one had corrected it, though a few other answers mentioned buggy behavior. After spending a few hours trying to figure out why my code was broken, I discovered the spec was changed.

An obsolete flag would have been a great option in this case.

But I don't think that's enough. It is important to be sure that there is an equally good or better answer that is up to date (maybe even a CW?) on questions with an accepted obsolete answer. Just flagging as "obsolete" but leaving no up-to-date answer in its place is asking for trouble, especially since these answers are often highly-linked to from outside and appear high in search engine results.

Another thing that needs to be considered is that these are frequently used as dupe targets. Perhaps there needs to be a warning to dupe-voters that the accepted answer is obsolete.

Why an obsolete flag and not one of the other solutions? The community probably won't ever "prefer" that answer - even with scoring changes (and meta-effect ;), I doubt my up-to-date answer would ever surpass the accepted one. The code isn't dangerous, it just doesn't work as expected. Downvotes aren't really warranted, because it's a good answer. Giving rep to a user who makes a substantial edit might be OK, but that puts A LOT more pressure on the reviewers to "Get it right." And then what happens if it gets rolled back and someone with the editing privileged makes a substantial edit?

So, for a question like this, a wiki-style obsolete banner confirmed by gold-badge tag users would be the best solution.

Until the flag is implemented, edits can certainly be approved, if:

  1. It's very concise, consisting of a brief synopsis of what's wrong and a link to the spec/docs that demonstrates things have changed.

  2. There is a link to the spec so reviewers/future visitors can see you're right. (See #1)

  3. The edit looks like a later edit and not part of the question.

These three points pretty much directly apply to the "obsolete" banner as well.

  • 2
    I like the idea of using a banner as a primary avenue. Programming for new flags is somewhat specific to SO (do MathSE answers become obsolete??), and new programming is expensive/prone to unanticipated consequences. As you, @hildred, and others have noted, the banner affords a prominent notice about the version dependence specifics.
    – hardmath
    Aug 5 '15 at 13:49

It's a real problem and often comes down to software versioning based on technology change. And as said, the outdated answer to the current software version is still valid for that particular version.

Based on this, the SharePoint Stack Exchange site has always included a version-specific tag - just because configuring Search Service is very different in SharePoint 2007 and SharePoint 2013.

Enter image description here

When questions are asked, members of the community often ask which version the question belongs to, and when the information is given, the version tag is added.

As a user it's perfectly all right to ask the same question targeting another version without the risk of getting marked as a duplicate. It's true as long as the answer is different, since not everything change between versions.

We know that this isn't optimal and that we're using tags in a way that we're not supposed to do. But it works and is appreciated. Legacy questions and answers are left untouched, and new version-specific questions are added, without interfering with each other.

One could also think of a version-tagging highlighting such as the bug, discussion and support tags on Meta.X.SE. Another suggestion would be to separate the tags from version tags, even if it might complicate things.

  • 3
    This is a quick solution that will cover a lot of things that can be versioned, but sometimes it doesn't make sense, something might work in Spring 3.4.3 but be removed in spring 3.4.4, so now you have a Spring 3 tag and a Spring 4 tag, creating a tag per minor version is a bit much. Aug 1 '15 at 9:09
  • 3
    "It's true as long as the answer is different," - the problem is the OP doesn't know that when dupes are suggested to them, and reviewers might not know that either, the ones most likely to know are the ones able to answer. This is just an observation, not any type of statement against your answer, and I don't have a fix for it... just stating that it's not always obvious to everyone when a dupe is not a dupe because the older question was specific to a particular version. Aug 1 '15 at 14:11
  • @AaronBertrand True! This isn't an ultimate fix everything solution, but often we (on SP.SE) have to ask for version before answering a question. That means the version is present before anyone can post an answer to the question. I think this is possible on other networks as well, if there is a difference between versions. But if there are major difference between minor versions, Jan, our system fails... Aug 1 '15 at 14:41
  • 2
    @Benny With SQL Server in particular, the same kind of things can happen - we ask for version all the time, because there are often better solutions on newer versions. The problem is that a lot of people still use the old, deprecated methods even on the newest versions, and continue encouraging those ancient approaches in answers (without asking for version, or ignoring that version was specified, or - in most cases - not knowing that there is a better way in newer versions). Aug 1 '15 at 15:15
  • I see a few problems with using tags that way. You are using the tags to communicate information about the answer not the question. If there is one answer applying to version 2007 and 2010 and another applying to version 2010 and 2013, how should the question be tagged? And if the question had two answers which worked equally well on version 2013, but only one of them will keep working with version 2016, how would you communicate that with tags?
    – kasperd
    Aug 1 '15 at 17:22
  • I know the workaround system we use in SP.SE has it's flaws. One of those is what you mention - questions that could belong to both 2010, 2013 and 2016. That's why the version tag is explicit, and there can only be one. If the question spans over several versions - no tag is allowed. Aug 2 '15 at 4:53
  • My thoughts on tagging appear under a nearby answer: meta.stackexchange.com/questions/261817/… Aug 2 '15 at 5:04
  • 1
    @JanVladimirMostert If something works in 3.4.3 but is removed in 3.4.4, then someone is almost certainly doing something wrong. Such minor version number increments should indicate bug fixes, not removal of features. If the answer relies on undocumented behavior (whether a bug or an unintentional/undocumented feature) then it should, at the very least, state exactly which version that answer works in. That way, the answer will never become obsolete, the same way a question about Windows 2000 isn't obsolete just because we now have Windows 10, a decade and half later.
    – user
    Aug 5 '15 at 12:55
  • Here's an example where upgrading from version 1.5 to 1.6 fixed a problem for me: stackoverflow.com/questions/22608289/… For Java, jumping from 1.6 to 1.7 to 1.8 is considered a major jump so using normal tags is fine, whereas Python jumping from 2.6 to 2.7 is considered a minor jump with a gazillion versions between 2.6 and 2.7 and if you have a Python2.1, Python2.2, Python2.3, 2.4, 2.5, 2.6 etc, you're going to clutter the tags. Rather have a dropdown on the tags where you can vote, works for me in version xyz, doesn't work for me in .. Aug 5 '15 at 18:22

Suggestion: emphasis on simplicity, and what's known to work well

Encourage use, greater and more widespread use, of existing approaches:

  • bounties
  • chat
  • comments that invite people to chat

– and so on.

Please, do not lose sight of the essence of Stack Exchange, in particular:

  • no distractions.

With that essence in mind, I should view any/all of the following labels/badges as distracting; as unnecessary and potentially troublesome:

  • old
  • outdated
  • obsolete
  • potentially dangerous
  • forcibly demoted (exempt from a top spot)
  • no longer acceptable (a notional de-emphasis to morph the check mark (tick) into something less than that mark in some situations).

Natural difficulties with masses of ideas

In the opening question I see two key points:

a good hard look


Lots of things have been suggested over the years

I find intrinsic goodness in Stack Exchange, and in its meta areas, so be realistic. It's hard, simply hard work to summarise and extract from so many years' worth of diverse suggestions. I offer the following paraphrase:

  • excesses of possible approaches may lead to difficulties that are unexpected and/or unnecessary.

That's a contentious phrase, and it may be construed as oversimplification, but it's heartfelt. In plain English:

  • don't make lives unnecessarily difficult for yourselves.

In this question, and in related questions, there's much to take in. Gut feeling: too much. For what it's worth I up-voted https://meta.stackoverflow.com/a/272691/38108 because almost without exception:

  • bounties demonstrate goodwill and are immediately, universally understandable.

On obsolescence

Here be dragons!

Perceptions of obsolescence vary wildly. An author or co-author may, understandably, take offence at their work being branded as obsolete.

(Compare with the tradition of offering bounties. I suspect that it's extremely rare for a bounty to cause personal offence.)

Beyond Stack Exchange

I think first of two 2015 topics in the Site and Forum Feedback area of MacRumors Forums:

Hint: you might plough through both topics, but you'll not get the full picture. Contentious, difficult contributions have been made invisible. From what's visible, in a nutshell:

  • there'll be no lasting/worthwhile agreement on how words such as old, older, classic and retired should be understood.

Within Stack Exchange

I think first of a 2011 question on Ask Different, Listing all Openmeta Tags and associated Files. It seems that part of the accepted answer is outdated due to a surprising behaviour in a more recent version of a third-party product; complicated by the relatively loose tagging of the question (nothing to indicate that it should not apply to Mac OS X v10.9 (Mavericks)); potentially complicated by some readers who will assume, wrongly, that because Mavericks tagging was based on OpenMeta, the answers will apply to Mavericks; and so on.

A person might strictly, logically, mark that accepted answer as obsolete but doing so would surely cause offence.

There be dragons!

  • 4
    I agree. "Obsolete" isn't just contentious--it's not even objective. When is an API obsolete? When the next version is stable? When the next version is widely adopted? When the next version is released? Aug 3 '15 at 21:22

I would go with Introduce an "Obsolete Answer" vote


Recently I came across an issue over this which I later discussed on meta. Here according to the user all other answers were outdated just because another user used a FlexBox solution which was not that well supported even when the user answered the FlexBox solution.

In such cases, these users will misuse the Obsolete Answer vote. How we can exactly define what's obsolete? Some methods which are deprecated officially are still used by many many people out there. Forcing such answers to change the solutions to the latest might be a wrong thing. I can relate this to mysql_* PHP questions where PHP users use to comment on each question or answer, but it is really not that easy to change the code base if it's huge and for such people, referring to altered answers from mysql_* to mysqli_* or PDO might not be a right way to go for.

So the question is what to do?

I would say we can use something like a flag option that a user can flag the post as outdated. In there he can propose a solution to the OP by either explaining him what was wrong, even the code if it was outdated. Here we can let the OP decide whether he wants to consider the suggestion by X user or he can alter the answer all by himself. We can also show a warning on the top of the post saying that your answer was flagged as outdated and we can keep a separate tab in the profile to maintain that queue for outdated post.

Coming to what if the user is no more active. In such cases if these flags increase in number on that particular post then we can move them to a review queue where other users will review and drop in their suggestions before the answer getting accepted by the community.

Note that such a workflow will only work well if we allow only professionals to suggest and accept the review queue. Also it will be better if we only allow users to flag the posts as outdated having a certain amount of score in that technology tag.

So a user having a couple of points in the PHP tag shouldn't be allowed to flag any post as outdated whereas users having, say, more than 500 points should be allowed to flag whereas users with gold badges in that tag should be allowed to review and accept the edited posts if the user is inactive and the obsolete score reaches some threshold.

  • Some methods which are deprecated officially are still used by many many people out there. - Which is why I'm suggesting we keep this for serious cases. Having 5 people close stuff instead of one has kept a lot of illegitimate closures away. I believe the same can go for "flag as obsolete"/"vote as obsolete"/whatever.
    – M.A.R.
    Aug 3 '15 at 8:03

I recommend and practice a simple two step method:

  • edit the existing post to include a version or date qualifier. This is less likely to be seen as changing the intent and preserves the answer for those working with old versions.

    • Adding a flag as suggested by others would complement this as it would allow downsorting obsolete answers.
  • add a new answer either as an addendum to the other answer with attribution and a link or a completely new answer which includes version information.

This gives you the best of both worlds.

  • 3
    This is great, except the fact that in smaller or bigger SEs, for different reasons, some important edits go unnoticed. Furthermore, getting people to edit is a chore itself. Also, most people don't scroll far down to see the updated answer on the bottom of page 3 of answers. So, you need to propose something that'll help with the sorting.
    – M.A.R.
    Jul 31 '15 at 22:36
  • 3
    If someone can't figure out that they need to scroll down farther when they are using version five when there is a bold line at the top of the answer that the answer applies to version three because stuff changed in version four, then I really don't know what to say, but if we use an outdated flag it would be reasonable for the obsolete answers to be downsorted. my proposal for comparative voting would also help, but no one liked that.
    – hildred
    Aug 1 '15 at 0:01
  • This idea may be useful as is on technical stacks, but on stacks such as cooking, parenting, cycling, etc, it would need to be broadened.
    – andy256
    Aug 1 '15 at 10:24
  • @andy256, The advice about bacteria on wood cutting boards seems to change every time someone does a study (which seems to be both technical and suited to dates for tracking recommendations), so I'm not particularly concerned. either it changes rapidly (release of a study, change of law or regulation) and would be suitable for marking with a date or version, or it does not change which is not a problem, or it changes slowly which is a different problem (new answers are discriminated against which I think should be solved by a different sorting algorithm like comparative voting or date biased).
    – hildred
    Aug 1 '15 at 19:44
  • I do like the simplicity. However it's sometimes inappropriate, or impossible, to edit as suggested in the first of hildred's two points. chat.stackexchange.com/transcript/message/23022138#23022138 as a recent example of refraining from edition of an answer to add version and/or date information. Aug 2 '15 at 5:02
  • Suggesting a few critical improvements: 1)the disclaimer must be prominent - at the top of the post and highlighted (quotation block perhaps); 2) it shall tell where the old answer stops working (it's not always a version) rather than where it works; 3)it must not link to another answer directly but should probably tell the reason, concisely (a footnote or a link to a comment perhaps if more than a single phrase is needed). E.g.: Irrelevant since vX.XX due to new_function(). Feb 7 '16 at 2:45

I find it helpful to start with a concrete (though hypothetical) situation. There is currently no predefined way to copy all the values in a Lua table. This turns out to be a problem that Lua beginners face fairly often. As a result, my question on the topic became Famous and I got a pleasing variety of answers. What should happen if a deepcopy() function were added to core Lua?

I'm not sure the right move is editing the existing answers, which continue to answer the question in a correct, if old-fashioned, way. But I could see someone either:

  1. answering with the benefit of new information, or
  2. editing my question to show that the answers were provided before the date the function was introduced.

The second option has the additional attraction that the person discovering the outdated question could ask and answer the question for the current release of Lua. As long as the two questions are linked in the sidebar (or the original question closed as a duplicate of the more general question), I don't see many downsides to this approach. Maybe some lost PageRank on posts that have been linked to all over the internet?

Let's push a little harder on the first option. When only a few people have voted on outdated answer, it's clearly a better option to write a new answer since that will provide reputation to the person who found the problem and get the correct answer to the top of the stack without much effort. Outdated answers just aren't a problem if the top-voted answer is less than a certain threshold. Exactly how many upvotes is too many, I'm not sure, but we are certainly well into the long tail.

Even so, there are thousands of questions with lots of upvoted answers that could potentially become out of date. I've tried various ways to detect outdated answers. My most recent attempt looks at answers with anonymous negative feedback a year after the most last edit. For the most part, I don't see answers that are obsolete, but rather controversial. I'd guess that most of the answers that seem incorrect are less out-of-date and more differences of opinion. It's difficult to think of a way to encourage productive edits to obsolete answers without also encouraging unproductive edits to opinionated answers.

Ultimately, I think the solution is to think of the problem being obsolete questions rather than obsolete answers. Often technology changes in order to make things that once were difficult and annoying (i.e., good questions on Stack Overflow) easy and sensible (i.e., better explained in the documentation). In these cases an edit to the question noting that the problem is no longer a problem for people using the latest release seems more expedient than trying to change each of the answers or figure out how to manipulate the voting to get the updated answer to the top of the list.

  • 1
    Thanks for posting and running the stats, we'll have to wait a bit though since SEDE is offline at the moment. Also a great 'thank you' for looking at the existing SE mechanism and its still largely untapped potential. Aug 4 '15 at 22:18
  • 1
    @DeerHunter: Try now. I've run out of ideas for trying to find obsolete answers. I honestly have no idea how big a problem we have. Aug 5 '15 at 0:05
  • @JonEricson What's your take on the new answers being pushed down below an "older" accepted answer? If the older answer gradually gets enough DVs it could end up with a negative score but still be pinned? Do you think we should reconsider to discuss the possibility of no longer pinning an accepted answer to the top? Or even consider a new sorting tab by "newest" instead of just votes, active, and oldest?
    – Taryn StaffMod
    Aug 5 '15 at 16:57
  • 2
    @bluefeet: I'm not a huge fan of pinning accepted answers in the first place. I do enjoy being able to prioritize answers to my own questions (see the Lua example in this answer), but it feels out of step with our goal to become the Wikipedia of long tail questions. It reminds me a bit of how the creator of a subreddit becomes "moderator for life" over it. At the very least, I'd like to see the priority age away over time. Aug 5 '15 at 22:38

Couple of ideas:

  • Mark as a Legacy answer - far more positive than Obsolete
  • Provide badges for many legacy answers to add a positive spin
  • Gamify editing by creating a separate edit rep with separate rules ie. +2 per edit upvote, Max +10 per upvote to post etc.
  • Gamify certain types of clean-up flagging with separate flagging rep and allow voting on raised flags, with a maximum rep cap
  • Create a tag version chaining system, and append tag version to bottom of legacy post, with latest tag version added to question when answer for that version is added
  • Please offer an example of something that might be marked as legacy. Aug 2 '15 at 4:45
  • @GrahamPerrin second sentence of the question states: "We don’t want to lose or delete these answers, due to legacy applications that still need the code..."
    – Dom
    Aug 2 '15 at 5:12
  • Sorry; I meant, for you to offer an example of an answer to something that's not in meta that you might mark as legacy. Aug 2 '15 at 7:28

How to Encourage More Up-to-Date Answers (A Modest Proposal)

I see that the topic of aging or obsolete posts has been on the discussion list for quite a while now. And I understand why it is hard. Perhaps looking at the problem from a slightly different angle might help.

A suggestion: Instead of trying to purge the aging answers the site should provide a mechanism to support redirection for a Version Update. Why? Remember that the answer was once accepted and that downvoting, obsoleting, or deleting are all potentially contentious actions.

  • Downvotes for aging posts will not resolve the problem, as has been discussed in detail already.
  • Obsoleting an answer is not particularly helpful. "Obsolete" versions of software can live for a long time and people may need the help for years. (How many of you are still supporting "unsupported" software? Me, too.)
  • Version Update, however, can be used to provide a strong indication that a more up-to-date answer may be the better answer for someone who is searching for more current versions.

Proposal: 3 simple steps

  1. A knowledgeable person creates a new question on the subject and nominates it as a Version Update for a particular question. Then he can self-answer the question with his new version of up-to-date answer. (If he is not able to self-answer the question, then the question might provide bait for someone else to answer it.) Since this is likely a version specific issue then proper tagging is essential and perhaps should include a Version Update tag.

  2. Members who pass some bar (points, posts on the topic, etc.) can vote up or down on the proposed Version Update to determine whether it should be a redirect. (If more control is needed, then perhaps a Moderator will finalize the redirection entry.) Then the redirection comes to life.

  3. At that time: (1) the link to the new answer would appear as a header on the previous answer, and (2) the link to the previous answer would appear as a header on the new answer.

I picture these links as looking similar to the This question already has an answer here: box for duplicate questions. Perhaps entitled something like:

This question has an answer for a newer version:
Newer link appears beneath the caption

This question has an answer for a previous version:
Previous link appears beneath the caption

This honors the work done in the past, but still provides a public venue for providing updates.

  • 2
    This! The fact that obsolete answers may still be required by many (because the industry is slow to adapt) means that multiple versions need to cohabit, and mixing up answers about various versions in a single question is asking for trouble: none is "better" than the other, it breaks the model. Therefore, one question per version, with a link to all other versions is a start; and then, I even wonder whether links should not automatically be redirected to the "newest" question so that traffic is driven to the most up-to-date answer (the obsolete one still being available anyway). Aug 4 '15 at 9:35
  • There seem to be two possible variants of this: a version update of an answer, or a version update of an entire question. It looks like you are proposing the former. Have you considered the latter? The latter circumvents a lot of issues with manipulating answer sorting order, etc. and seems like it might be a lot cleaner, while still being adequate.
    – D.W.
    Aug 5 '15 at 1:07
  • 1
    I agree that a version update of the question is what I intended, and I have now circled back to take a better stab at the Question/Answer approach. This way the Q/A behavior remains in place just like all the other questions and answers.
    – RLF
    Aug 5 '15 at 2:09
  • Disagree: 1)linking to a single answer creates monopoly rather than letting other answers compete; 2)version is not always a cutting point; there are areas where there's no such thing as a single version number at all (decentralized projects or where there are many build options that greatly affect functionality). Feb 7 '16 at 2:30

I would like to see the simple option to sort by date in both directions.

Currently, I can sort by "oldest", but then I have to scroll all the way to the bottom of the posts to be able to see the newest answer - which may not even have any upvotes yet. Sorting by newest would allow people to add posts for updated info without messing up anything related to votes.

My experience is that the "accepted" answer isn't always the correct or best answer, anyway. Perhaps the OP accepted the answer before actually trying it out, or perhaps something looked like it should work, so they accepted it, or ... Regardless, as a result, sorting answers by upvotes AND/OR date, especially for those legacy questions where updated answers HAVE been posted, is invaluable.

As a note, I've been hit with the "duplicate question" deal, even when the original question and all its related answers were several years old. I attempted to ask about functionality in Visual Studio 2012 (which was stated in my post) and got shut down because the same question was asked and answered in something like 2007. That's just crazy.

I did want to attach a version-specific tag to the post, but none existed, and I didn't have enough rep to create one. So, when you consider version-specific tags, also consider how tags will be added for any new versions.

  • 1
    In a case like that, putting version info specifically in the question (but not in a tag) and potentially commenting to ask some helpful 1.5ker to add the tag should work quite well. Aug 2 '15 at 2:26

For the specific case of an accepted answer becoming obsolete I've two suggestions that basically boil down to "don't treat the accepted answer as special".

  1. As I mentioned in the comments and as has been suggested on MSO - Order highest upvoted answer before accepted answer - it would be helpful to just show the accepted answer in its "natural" vote order. Then if it drops off the top spot there's an incentive for the poster to update it. Also if it can be toppled from its spot underneath the question there's an incentive for others to down-vote it should it become completely outdated.

  2. Don't prevent regular users deleting an answer just because it's accepted. Again, this only helps if it's completely obsolete.

  • I was thinking about mentioning this; I wonder why the change has not already been made.
    – Honest Abe
    Aug 2 '15 at 18:39
  • @HonestAbe emphasizing accepted answer denotes a social contract (as Atwood calls it) and serves as convenient differentiator between SE "Question and Answer" model and that of typical Internet forums. Obscuring it carries a risk of (further) eroding Q&A model. To avoid misunderstanding, I am not really against making such a change, but merely explain why we better thread carefully here
    – gnat
    Aug 4 '15 at 19:43

I think we have two problems:

  1. Identifying such old posts
  2. Get someone who updates it

I also think for identifying those posts anybody with an acceptiable score in at least one of the tags, maybe 1k, can raise a flag that this post is "old and needs an update". Then the post goes to a queue where this flag can be reviewed by higher rep people - again of the included tags. So people who are obvious familiar with the content can identify these posts.

After it is reviewed we need to get attention to it and try to find someone to edit it: I'm a low-rep user, so I don't know if reputation is interesting for someone with higher rep, but putting out a kind of bounty (maybe not reputation but something other - sorry, I don't have an idea in which higher reputation people are interesting in) to the person who edits it. To verify the edit it must be reviewed again even if you have the edit instantly priviledge.

I think this can be a way to motivate people to edit and not only to flag the questions.

Aside from that, I also like the idea to give people the possibility to also tag their answers with version-tags. I saw several very good answers on Stack Overflow which were written like: In version X you can solve your problem with this code, in version Y it's like that. Tags may help here - but they have to be optional as not every answer is versionable. For sorting I could imagine that answers with a newer version tag are listed over older versions if they have positiv score, like this:

  1. Answer 3, 2 Votes, Version X
  2. Answer 1, 200 Votes, Version W
  3. Answer 2, 100 Votes, Version W
  4. Answer 4, -1 Votes, Version Z
  • I would use the relevant chat room to write things such as "old and needs an update"; and I might invite passers-by to place a bounty in the required area. Aug 2 '15 at 4:48

Add tags that show a versioning and remove the general tags without versioning. For example, remove the general tag sql-server and add maybe a tag, sql-server2005. Therefore the older and obsolete answers will only show up if you look explicitly for them.

  • In theory, tagging can be great. In practice, tagging can be a chore; is limited; and can be misunderstood. Aug 2 '15 at 4:51
  • More on tagging in nearby meta.stackexchange.com/a/261838/166799 Aug 2 '15 at 5:05
  • 1
    This would have to be a huge, network-wide issue in order to deal with the hundreds of thousands of non-version-tagged questions. It might run into issues, too, when you deal with questions that apply to multiple versions, but not all (or that apply to all versions before a certain one). There is a five-tag limit to questions, as well.
    – TylerH
    Aug 2 '15 at 21:54

I think:

  1. The upvote effectiveness for ranking should be affected by date and/or version.
  2. People should be able to filter questions by date and/or version.

This means that people can quickly ignore answers that are too old/too new to answer their problem. (I can't think of very many other benefits though)

I also think version tags should be grouped under the general software tag, i.e. when you select .NET-2.0 the tag should look something like [.NET[2.0]]. This also would mean that the question would appear in a search for both [.NET] and [.NET-2.0].

The answers could provide versions as something like answer tags.

I also think that answer versioning is a good idea, with a timeline/arrows to switch between different versions.

  1. Enter Useful information with the answer

    The user should know the context of the answer I will take as an example an iOS answer the user should know for which iOS version the answer works and in which version of Xcode it works , and also in which language Objective-C or Swift and which language version Swift 1.2 , Swift 2 ... And these information should be searchable, so the user can search for question that have answers about something that works for iOS 8 for example...

    To do this we should force the user to enter while answering in which iOS, Xcode, language and language version his solution works.

  2. New "Updator" badges for updating obsolete answers

  3. Deemphasise the accept mark if there's an answer the community strongly prefers

  4. Show an "obsolete" notice like "duplicate" and propose an up-to-date similar question.


How do we encourage edits to obsolete/out of date answers?

You could use reputation points.

Allow for users with over 2000 reputation to choose to have their edit reviewed.

Each reviewer can either vote against the edit, or grant 1-3 points.
That way minor edits would get 3 points, and noteworthy edits might get 9.

  • 1
    Please explicitly respond to the points raised in the section of the question with the header "Why, you ask, do I see it as a stumbling block?" Aug 2 '15 at 0:05
  • @Mr.Bultitude In regard to fixing code that is using some deprecated function or feature, I think stating that fact in the edit summary would be enough. In the situation where the whole answer is obsolete, I think writing a new answer is more appropriate.
    – Honest Abe
    Aug 2 '15 at 0:46

I really like the "Obsolete Answer" option, but what I would like even more is a slightly different option that would supersede the need for marking accepted answers as obsolete:

Introduce a new reputation perk that enables community voting on changing the accepted answer.

I first thought of this idea several months ago:

Let users with rep over X (15k, or 20k, or 50k, or maybe just with a score of >2k in that tag, whatever) accept answers to questions that have no accepted answer AND are over six months old AND when the user has not answered that question.

While that is very similar, it's not exactly a good fit for what this question is asking. Morphing that idea into one that better fits this issue would look something like this:

Let users with rep over X (10k, or 20k, or maybe with a score of >= 500 in that tag) cast votes to change the accepted answer to questions that have had no accept activity (e.g. the asker has not accepted an answer or switched acceptance to a new answer, etc.) in the past six months AND where the voting users are not voting on their own answer (voting on community wiki answers that used to be yours would be OK).

At least five/six votes up to around ten votes would be sufficient, I think. No increase in vote weight with increased reputation or tag score, however. If Stack Exchange wanted to get really ambitious, they could even create a new review queue for "accepted answer votes".

I chose a six month lower limit because of the emphasis on old technologies in the question here, otherwise it could probably be something like three months.

  • Any of the (currently 3) downvoters want to bother describing in words their disagreement with this sorely needed suggestion?
    – TylerH
    Aug 3 '15 at 14:28
  • As the first person to downvote this, this was my perspective: Accepting the answer only makes sense for the original poster. Your suggestion is essentially nothing more that providing a super vote for the elites, adding an additional and unneeded factor into an already confused situation.
    – corsair992
    Aug 5 '15 at 19:20
  • @corsair992 In a normal context, sure. In the context of this question, obsolete answers are clearly problematic, and need addressing. I agree it's not a perfect solution, though (It's a better suggestion for when there is no accepted answer). As far as "adding a super vote for the elites, adding an additional and unneeded factor", that's exactly what every other answer advocating for the "obsolete answer" option here is suggesting, so I hope you downvoted them as well :-)
    – TylerH
    Aug 5 '15 at 19:29
  • If we are to implement a specific technical solution for "obsolete" answers at all, then confusing it with "accepted" answers is not the way to do it in my opinion. I don't necessarily agree that we need a new solution here unless there is an extreme problem to address (as there are already many ways to deal with this), but I am not so strongly against it as to downvote those solutions.
    – corsair992
    Aug 5 '15 at 19:55
  • An underlying problem might be how prominent the "accepted" checkmark is, and how its prominence confuses users as to what it means. Technically it just means "this post provided the best solution to the asker's question or problem, in their own opinion", while in practice it means "this is the best answer in the asker's opinion", but, importantly, to many users it just means "this is the best answer". Personally I think the checkmark should be smaller, and accepted answers should not be at the top of the list for all sort options, but that's quite a separate topic.
    – TylerH
    Aug 5 '15 at 20:17
  • Painful experience compels me to agree that we need to stop forcing accepted answers to be displayed at the top in all situations. In principle the asker should be in a position to evaluate the best answer, and stay around to accept any better answers posted later, but in practice both of these things don't seem to be happening. Perhaps the check mark should just be evaluated as a vote by the asker with an appropriate weight, and sorted accordingly. I seem to remember a few posts on the subject on the SO Meta from a while back.
    – corsair992
    Aug 5 '15 at 20:38

PLEASE Save Obsolete but Useful Answers, Using the Current System

  1. Moderate and Edit Questions, which have obsolete answers, to reflect the version the old answer applies to.
  2. Create another question, with the qualified by the version, period, etc, and link it as a closely related question.

Further, implement "Tabs" for VERY closely related questions

Old answers are very helpful in view of deprecated versions--especially in technology, and mechanical engineering

Using the current tag framework, Tabs could be implemented:

Stack could implement Parallel Questions/Answers for different versions. These closely related questions/answers would be displayed in a different "TAB" at the top of the page.

There is a real problem in questions like: In C++ 98, How do you simulate Constructor Delegation? Rather--Questions could be linked, with tabs: In C++, how do you implement Constructor Delegation? And then have the community define a "Tab" for each individual form of the question, In C++ 98, 11, 14 ... etc.


  1. C++ Constructor Delegation, How To? ... Tabs: C++ 98, C++ 11, C++ 14.
  2. What is the meaning of this X word? ... Tabs: M-W, Oxford, etc.
  3. How to Implement MVC in .NET? ... Tabs: v3.0, v4.0, etc.
  4. What classical psychoanalytical approaches are recommended for the treatment of PTSD in Children? ... Tabs: Jung, Freud, (intentional bad example).


I can foresee that some people try to post alternative answers in different tabs--but this risk is present in the other solutions as well.

There would have to be a "Hard/Fast" rule that tabs can only represent "versions", or Answers in view of a Different referential authority.


New answers to old questions hardly ever get upvotes.

Most of my answers are to old questions. I write them when I face a problem whose solution wasn't on SO. (I don't just "hang out" on SO, but I try to give back.) Sometimes I write the question as well, but more often one exists.

Typically, the existing answers are not "obsolete." More often, there simply wasn't a good answer. Either way, I write my own.

I think writing a new answer is better than editing someone else's. I don't rewrite fresh answers, so why should I rewrite old ones? That's disrespectful to their author--putting my words under their name. What if I'm wrong?

Writing an answer to a question even a day old hardly ever gives reputation. No one votes on answers to old question. It doesn't faze me, but it feels a bit unfair. My profile says I've "reached" three quarters of a million people, but just 0.6% left a vote.

I think key is to get the lurkers coming from Google to vote on answers. First step may be to let unregistered visitors vote. Second step is to make them feel that their vote "counts." A vote counts when the number of existing votes is low, and that can be achieved for popular questions by letting old votes decay. Counting votes in a moving temporal window gets rid of "obsolete" answers as well. Reputation should be rewarded for voting, which can be dangerous, but surely we can profile user actions to detect thoughtless serial voting. We can reward votes on old questions more than on new ones. (And, if all else fails, we can reward new answers for old questions more as well.) Finally, we should allow "thank you" comments since everyone who contributes likes to feel that they're actually helping people.


My addition: I wouldn't mind being able to sign up to be pinged when one of my old answers gets flagged as "Legacy" or "Obsolete." This could be tacked onto some of the other ideas here such as versioning.

Just a little icing on the cake but the person best able to decide whether an answer remains relevant is the original author.


3 parts to this answer

1) mark as legacy /icon on the question its self. denote this as an outdated system. may need review. [ex windows 95 ini edits vs xp registry edits. xp registry edits may work on win7 where 95 no way... ] once its sure that no current/new systems use that setting mark the question as legacy with a icon so readers know it will NOT work on current things and possibly make a notation win95-98 type blurb available to state what it will or will not work for and possibly a comment stating the reason for the legacy mark.

2)valid but dangerous gets an icon/comment by the answer. ie edit hklm\current user\software if its a 64bit system you may break things if you dont go to hklm current user\syswow64 or what ever the key is for 32bit apps.

3) update for new information/system icon. if say the question is a year or 2 old and some one wants to put a new answer there is now a U icon for the new updates if nothing else hovering left side so if they click on it the page scrolls/ highlights the updated answers. requirements for the U tag are a preface sentance stating why this is added.

Bios cant find my hard drive why?

[Update: Uefi bios has additional and different options released 2012...]

Rep and such required to do these is hard to say. and voting to have a question reviewed for legacy which should take a bit of rep should also be available.

  1. (to start with) Only worry about questions with >= 3 views per day, older than 1 year, and has accepted answer.
  2. Let anyone with >= 5000 rep cast an obsolete vote on the question which will mean that all the answers are obsolete now. Be clear that this vote is not for "not enough good answers on this question, or this question could use a better answer". That it's only for questions where the accepted answer is obsolete and there are no other answers that provide the new correct way of solving the issue.
  3. Create an obsolete answer review queue. You need at least a bronze badge in the most popular tag of the question to be able to review it. Action choices in the queue:

    • not obsolete
    • answer
    • skip

    Keep the review active in the queue until 2 people provide an answer, gets enough not obsolete votes (3-5?), or until 30 days (or so) has passed. Make a post history entry on the question with the review link, so easy to search SEDE for questions that were voted as obsolete but didn't get any new answers for that 30 days. So a second chance for people to find and answer even after it exits the queue.

Why no edit option? Because reputation is the fuel that drives good content. You don't get reputation for editing (except for measly +2 for suggested edits, but anyone with a bronze badge likely has over 2k anyway). I really don't think encouraging edits is the way to go since I think it will definitely be not as successful as encouraging new up-to-date answers where users can post under their name as their answer and reap the benefits.

But why not allow people who know the new correct way to answer in the review? Because this is a privilege queue where you need at least 1 bronze badge to be able to enter. After the review is completed provide a way for people to search through SEDE and regular search to find questions that made it to the obsolete queue and weren't completed as "not obsolete".

In conclusion, I really don't think encouraging edits in this manner is the way to go. Since really, reviewers of the suggested edits queue have long been advised to be wary of approving edits that change/add/remove code. I'd rather a new answer be placed than all the content-specific edits going on and the haywire it has the possibility of causing. Also, reputation incentive is needed to keep the ship sailing. Not sure about a similar review badge for this queue, I think if there is a badge for this queue it should be different, but any badge(s) can be sorted out later.

  • 1
    I understand the bronze badge restriction - to prevent those that aren't as knowledgeable in the technology from saying an answer is (or isn't obsolete) when the opposite is true. However, I think it also makes it harder to get an updated answer. A user without a bronze badge can provide an updated answer.
    – Andy
    Aug 26 '15 at 23:37

Do NOT touch existing valid answers. They're not "legacy", nor "obsolete" - they're perfectly valid answer for technology that author asked about at the time. When I for some reason have to work with same old technology and want answer to same question, I do not want to find dozens of answers about how to do it in shiny new way that are invalid from my point of view because they simply don't work at all in old environment.

If you want to provide a new answer without mixing it with old ones, ask a new question specifically targeted in title or text at new version of language/standard/software/whatever and provide answer with new way there.

Optionally you can also edit an old question to mention that there are better answers for version later than X.Y.Z in other Q/A.


"How do we encourage edits to obsolete/out of date answers?"

[Disclaimer: Actually read all the Answers and most of the comments - whew]

The problem has likely come about from a combination of The Rules, human nature, the way the site is setup (page layout), rewards (psychological, numerical, perceptual, expediency), etc. - this is evident in many of the answers.

Clicking on "Questions" takes one to the main Feed where the (under) Banner goes: "All Questions ... Newest - Featured - etc.", but there's not an "Oldest" where people could go to be nostalgic, to find questions known to have newer answers, or to hunt for things to suggest for deletion.

We promote 'new questions' as better (or more interesting) by not only having a link to click for them but likely the belief that answering the new questions correctly and quickly:

Rewards the answerer with UpVotes (with nothing to compare to),

Provides the possibility of being The Chosen One (usually viewed as good, sometimes an insult - to suggest that you are special),

Self-promotes one as the sharpest knife in the drawer.

In the interests of keeping this answer as short as possible while still being useful I propose:

A 'Legacy Review' Queue - While waiting for new questions or for one's question to be answered people could be encouraged to poke around where they could suggest an improvement and score a point. (Add "Oldest" to the end of the sub-banner Menu where "Newest" appears).

Maintenance Points - Approved edits, deletions, improvement, etc. score a point.

Trivial Edits - After accumulating enough reputation allow trivial edits (currently the minimum limit is a few characters must be changed). A good example is updating the question "How many Planets in our Solar System" or "How many Moons does Saturn have" - those answers don't require a big edit but they do need a current answer. It's also nice if someone reputable can fix a single annoying spelling mistake in an otherwise good question or answer.

The 'presentation order' for the answers should be:

Chosen answer first (as someone who's question is accepted is the one who accepts the answer). The Mods can still delete unwanted chosen answers, bumping it back down to unanswered.

Highest voted answer next.

Highest trending answer third - new correct answers should rise up over legacy unchosen and lessor-popular answers.

For some sub-sites a versioning system is appropriate, so we have the newest version's first three followed by the next version's first three, etc. followed by the rest, and the newest (low votes) untrended answers at the bottom.

Whatever is ultimately decided upon should be relatively simple, intuitive, cross-site implementable (with a provision for sub-sites to exempt themselves when applicable).

We can encourage edits to obsolete answers by providing a means to funnel people towards helping with them, to donate a half-hour during their visit, and score a few points, maybe a "Hundred hour Maintenance" Badge - it's not much but that's all we got.

Feedback Section

Some sites (EG: workplace.stackexchange.com/) would really benefit (and many comments indicate that people really appreciate it) if the person asking the question would come back and give us an update of exactly which advice they followed (sometimes it's not only the chosen answer) and how effective it was.

That helps everyone to know what worked best for their particular situation. The Feedback could also receive up or down votes to score a few more points. It also avoids the situation where a question is asked and answered but it's unclear if the asker came back to thank everyone when no answer is chosen.

I'll leave it there in the interests of not adding yet another lengthy answer to be read when there exists so many already and avoid contributing to the problem this question wishes to solve.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .